EP1207507A2

EP1207507A2 - Method and apparatus for reducing differential delay problems in audio communications systems with at least two transmitters

Info

Publication number: EP1207507A2
Application number: EP01308958A
Authority: EP
Inventors: Stephen Kenneth Cruickshank
Original assignee: NAT AIR TRAFFIC SERVICES Ltd; National Air Traffic Services Ltd
Current assignee: National Air Traffic Services Ltd
Priority date: 2000-10-23
Filing date: 2001-10-22
Publication date: 2002-05-22
Anticipated expiration: 2021-10-22
Also published as: ATE368915T1; ES2290099T3; EP1207507A3; GB2368492B; EP1207507B1; GB0025916D0; DE60129655T2; DE60129655D1; GB2368492A

Abstract

In an audio communications system with co-channel transmitters, as used for example on air traffic control, signals are pre-filtered prior to transmission in order to compensate for differential delays from transmitters to aircraft causing distortion. Either the phase or the amplitude of signals to be transmitted is varied by an amount which is a function of frequency or time.

Description

The present invention relates to audio communications systems in which at least two transmitters are able to transmit identical audio signals to a receiver. Systems of this type are sometimes used in the aviation industry for communication between air traffic control centres and aircraft. One such system is illustrated schematically in Figure 1. As shown in Figure 1, audio, ie, voice, signals from an air-traffic control centre 10 are conveyed over a network 11, to be discussed further below, to two radio frequency transmitters 12 and 13. The transmitters communicate with equipment on an aircraft 14 where the audio signals are reproduced. Figure 1 shows the particular example where an aircraft is approaching the South East of England from Boulogne on the coast of France. As the aircraft approaches England, it will receive air traffic control signals from the transmitters 12 and 13 at Winstone and Warlingham. The time taken for signals to reach the aircraft via network 11 and transmitter 12 is different from the time taken for signals to reach the aircraft via network 11 and transmitter 13. This difference is referred to as a "differential delay".
There are two factors which contribute to this delay. The first is due to the relative distances between the respective transmitters and the aircraft and the second is due to the different propagation times for the signals to the respective transmitters via the network 11. The effect of the differential delay is to cause a phase shift between the respective signals resulting in destructive interference at certain frequencies (as well as constructive interference at other frequencies), and hence significant attenuation (and amplification) and consequent distortion of the received signal in certain frequency bands. For most of its travel, the signal strength received by the aircraft will be unequal so that the signal from one transmitter swamps the other and the distortion is insignificant. However, the differential delay can cause problems where the receiver is receiving the same audio power from both transmitters. Equal signal strength is not necessarily at equal distances from each transmitter or radio station. The signal strength from each station may depend on many factors, some of these being:
Polar diagram of transmitter site antenna
Height (the ground wave and direct wave interact to vary the signal strength)
Land mass between the station and aircraft (especially long distance)
Polar diagram of aircraft antenna
Banking of aircraft.
In the example of Figure 1, transmitters 12, 13 are respectively 225km and 125km from the aircraft.
This causes a differential delay of: 100000m3 x 108 m/s (the speed of light in air) = 0.33 milliseconds The delay due to different propagation times over the network 11 is estimated to be 0.5 milliseconds, resulting in a total differential delay of 0.83 milliseconds.
The subjective of level of degradation at audio frequencies depends on the length of the differential delay, as different delays lead to attenuation of different frequencies. Ideally, the differential delay should be less than about 0.1ms. However, this is impractical to achieve because difference in propagation delays from transmitters to the aircraft can exceed this, even at locations where signals strengths from the transmitters are similar. Differential delays up to about 2ms tend to sound particularly bad because they cause broad nulls in the received audio spectrum. It has been found that higher delays are better for a radio operator because they result in many narrow audio nulls and peaks, a large proportion of the original audio power is present, and speech is readily intelligible. However, the delay must not be too great or there is an audible delay in the side-tone to the operator from the reference receiver, which is very confusing to the operator.
The problem can be better understood by considering the graphs of Figures 2 and 3. Figures 2(a) and 3(a) show variations with time of a typical voice signal received simultaneously from two transmitters separated by different differential delays of 0.20ms and 4ms respectively, whilst Figures 3(a) and 3(b) show power spectral density plots for the transmitted and received signals. These graphs have been obtained by computer simulation using real voice data.
It is clear from Figure 2(b) that a differential delay of 0.2ms causes severe distortion of the audio signal. Comparing the transmitted signal shown in dotted lines with the received signal shown in solid lines, it will be seen that there is a severe attenuation of signals above about 2,500 Hz. However, in Figure 3(b) there are many narrow nulls in the spectrum. (The audio is filtered above 3.4 kHz to limit the bandwidth). Thus little of the original audio information is lost, and the signal of Figure 3 is of acceptable quality.
The problem of differential delay has been addressed temporarily by adding a fixed delay in one of the links to the transmitters to try to ensure that the differential delay is always equal to 4ms. Incidentally, delays of this order have no appreciable effect on the r.f. carrier signals. However, this solution requires manual measurement of the differential delay and sometimes fails because of rerouting of the audio signal through the network 11. The network typically consists of digital telecommunications links, the operation of which may be controlled by a third party and out of the control of the air traffic control provider. Therefore, the air traffic control provider may not always be notified when re-routing of the network 11 occurs.
There is therefore a need for communications system which avoids the problems of differential delay even when signals travelling over the network are re-routed.
Ideally, the solution to this problem needs to satisfy all of the following requirements:
Audio quality independent of delay in the digital network feeding the transmitters.
Independent of the RICE (radio telephony interface control equipment) control and monitoring system (as this would be difficult and expensive to modify).
Compatible with the current analogue control and monitoring system.
Acceptable audio quality to aircraft receiving a single, dominant transmitter.
Acceptable audio quality to aircraft receiving two similar power signals from different transmitters (for both two and three transmitters).
Acceptable audio off-air side-tone quality to air traffic controller from receiver station.
Failure modes are important and should be considered.
Must require only minor changes to the existing communications infrastructure.
The present invention solves this problem by performing additional processing on the audio signal before it reaches the transmitters.
Thus, the present invention provides an audio communications system including an audio signal source, at least two transmitters for transmitting identical audio signals generated by the audio signal source to one or more receivers, and network means for conveying signals from the audio signal source to the transmitters, wherein the transmitters transmit to the receiver(s) modulated signals which are demodulated by the receiver(s) and at least one of the transmitters is provided with means for processing signals received from the network means prior to transmission whereby to change at least one parameter of the received signals by an amount which is a function of time and/or the frequency of the received signal. The invention also provides a transmitter for use in this system as described in claim 1, as well as a method of operating a communications system as described in claim 27.
In the preferred embodiment (system, transmitter or method) the signal processing means simply comprise a filter. The filter acts on the digital signal received over the network 11 prior to it being transmitted from transmitter 12 or 13. The parameter variation may be random but the preferred filter is one whose response varies periodically, for example sinusoidally, with time or with the frequency of the received signal. The filter may vary the phase or the amplitude of the signals. In the case of the system of the type illustrated in Figure 1, one or both of the transmitters may be provided with signal processing means. In the case where both transmitters are provided with signal processing means which vary the phase of the signal periodically with frequency, it is preferable that the phase change/frequency characteristic of the two signal processing means are 180° out of phase with each other. It should be noted however that the aim of the invention is to achieve at the receiver some of the audio frequencies being in-phase with each other although there will be some out of phase with each other.
In the preferred arrangement, the transmitters transmit amplitude modulated signals to the receiver(s), although the system may work equally well with other types of modulation such as frequency or pulse width modulation. Implementation of the invention need not require any additional hardware at transmitter stations. Many signal transmitters have suitable signal processors which simply require reprogramming to provide the necessary pre-filtering to put the invention into practice.
Embodiments of the invention will now be described by way of example only and with reference to the accompanying drawings in which:
Figure 1 is a schematic diagram of a conventional audio communications system in which an air traffic control centre communicates with an aircraft;
Figure 2(a) shows the variation in amplitude with time of an audio signal received from two transmitters at equal strength, delayed with respect to each other by 0.2ms;
Figure 2(b) shows the power spectral densities of transmitted and received signals corresponding to Figure 2(a);
Figures 3(a) and 3(b) are similar to Figures 2(a) and 2(b) showing the signals delayed by 4ms with respect to each other;
Figure 4 is a schematic diagram of an audio communications system according to the invention in which an air traffic control centre communicates with an aircraft;
Figure 5 is a graph of delay vs. time for a filter which delays signals randomly over time;
Figure 6 is a graph illustrating the response of a filter which varies phase sinusoidally with frequency;
Figure 7 shows impulse response, magnitude response and group delay for a filter imposing monotonic group delay variation over frequency;
Figures 8 (a) to (d) are typical plots of filter responses for filters which vary both phase and amplitude sinusoidally with frequency;
Figures 9 (a) and (b) are graphs of filter responses for filters which vary phase sinusoidally with frequency used in a two transmitter system;
Figures 10 (a) to (c) are graphs of filter responses for three filters varying phase only sinusoidally with frequency used in a three transmitter system;
Figure 11 (a) shows the variation in amplitude with time of an audio signal received from two transmitters at equal strength, delayed by 0.25 ms with respect to each other;
Figure 11 (b) shows the power spectral densities of transmitted and received signals corresponding to Figure 11(a) after prefiltering with the preferred filter for use in a two transmitter system according to the invention; and .
Figures 12(a) and (b) are similar to Figures 11(a) and (b) showing the signals delayed by 5 ms.
Referring now to Figure 4, items in this drawing corresponding to similar items in Figure 1 are identified by the same reference numerals. The system of Figure 4 is similar to Figure 1 except that at each transmitter 12, 13 a group delay filter 15 is present between transmitter radio station control equipment 16 and output transmitters 17.
It should be noted that although the two transmitters 17 operate on the same notional amplitude modulation frequency, in fact the frequencies are slightly offset so that radio station 12 operates at 125.005 MHZ and station 13 operates at 124.995 MHZ. The imprecision of the tuner of the aircraft is such that both signals are received and demodulated simultaneously.
As noted above, the degradation of the received signals when there is a differential delay between the two transmitters occurs because the two identical audio signals interfere with each other. At those frequencies for which the delay is equal to an odd number of half-wavelengths, the interference is destructive and therefore those frequencies for which the delay equals an even number of half-wavelengths interfere constructively, so that these frequencies are boosted. The effect of this is that the received signals can sound extremely muffled and/or distorted.
The intention behind pre-filtering the audio signals is that, by modifying the two signals in different ways, they become sufficiently different from each other that the amount of interference between them will be reduced considerably. At the same time, it is obviously important not to modify the signals too much or else they will become audibly distorted. This would be very unsatisfactory as the distortion would be noticeable even in the case of a single dominant transmitter. If the end result was comparable to that shown in Figure 3(b), this would be acceptable.
A number of different types of filter were investigated and their suitability tested using computer simulation. For each type of filter, both male and female voices were tested and a range of differential delays between 0 and 30 ms. In some cases a filter was applied at one transmitter only and in others the pre-filtering was applied at both transmitters. The filters which worked best applied a sinusoidal variation of a signal parameter (eg: phase, amplitude) with frequency, but others produced useful results. Examples of suitable filters are described below.
Some filters are defined in terms of their frequency response H(f). This relates the input and output signals of the filters in the frequency domain as follows: Y (f) = H (f) X (f) where Y(f) and X(f) are respectively the Fourier transforms of the output and input signals in the time domain y(t) and x(t). To allow for the possibility of two filters, two frequency responses are defined, namely H₁(f) and H₂(f).

Example 1

Sinusoidal delay variation over time for one transmitter.
This filter delays the input signals x(t) by an amount δ that varies sinusoidally with time. Thus, the output of filter y(t) is given by: y(t) = x(t-δ) where δ = Acos (2πt/T).
Parameters used for test purposes were A = 20 and T = 0.2s. Received signals using this arrangement were clearly intelligible although somewhat distorted.

Example 2

Random delay variation over time for one transmitter.
This filter is similar to the previous filter, except that δ is defined by picking a number of random values, n, between -0.5 and 0.5 and fitting a spline to the points to produce a smooth, but random curve. The amplitude of the curve was then scaled by a factor A. Different values of and n were tested. Figure 5 shows an example of the variation in delay over time, for A=20.

Example 3

Sinusoidal amplitude variation over time for both transmitters.
This filter applies a sinusoidal amplitude modulation to the input signal, so that the output signals from the 2 transmitters are given by: y 1(t) = x(t)[1+Asin(2πt/T)] y 2(t) = x(t)[1+Asin(2πt/T)] Again, different values of the parameters A and T were tested, eg A = 0.5, T = 0.01, 0.1 or 1s. Although signals were intelligible, the periodic changes in signal magnitude were clearly audible, particularly for the single transmitter case, and this reduced the clarity of the signal.

Example 4

Sinusoidal amplitude variation over frequency for one transmitter For this particular transmitter, the frequency response was given by: H 1(f) = 1 H 2(f) = 1 + A sin(2πNff'), where A is the magnitude of the amplitude variation and N_f is the number of cycles per unit frequency. In this and all subsequent equations, f' is the normalised frequency, defined as the frequency, f, divided by half the sampling frequency, i.e. f' = f Fs /2 , where (F_s = 8kHz).
Suitable parameters are A = 0.6 and N_f = 8. Note that the characteristics of the filter (as for all subsequent filters) do not vary over time.

Example 5

Sinusoidal phase variation over frequency for one transmitter.
In this proposed arrangement the two frequency responses are given by:
1st Transmitter H₁(f) = 1 (ie: no pre-filtering)
2nd Transmitter H₂(f) = exp (-jπP sin(2 πN _ff')

_f

Example 6

Monotonic group delay variation over frequency for one transmitter
As with the previous one, this filter modifies the phase of the signal in the frequency domain. However, rather than applying a periodic variation, it produces a monotonically varying phase (and therefore group delay) variation. The frequency response is: H1 (f)=1 H 2 (f) = exp(-jπPf'2), where, again, P is a constant defining the magnitude of the variation. Figure 7 shows the variation in group delay over frequency and the corresponding impulse response of the filter, with P = 128. Received signals are intelligible but this is not the best filter because it causes the received signals to sound very metallic. The single-transmitter output sounds particularly strange. In addition, it imposes a large overall delay on the signals at a higher frequencies, which would lead to a disconcerting amount of echo in the side-tone heard by the controllers. Nevertheless it may be useful for certain applications.

Example 7

Sinusoidal phase and amplitude variation over frequency for both transmitters. The frequency responses of these filters were specified initially by the following equations: H 1 (f) = (1 - A sin(2πNf f' + π / 2)) exp(-jπPsin(2πNff')) H 2 (f) = (1 + A sin(2πNf f' + π / 2)) exp(+jπPsin(2πNff')) From the frequency responses, the taps of the impulse responses were determined by taking the inverse Fourier transforms. A useful property of filters with periodic frequency responses is that the impulse response consists of only a few non-zero, equally spaced taps (this is a consequence of the fact that the impulse response and the frequency response are Fourier transforms of each other, and the FT of a periodic wave form is a series of discrete values). This means that the type of filter is very simple to implement and only requires a small amount of processing power. To simplify the filter, only the largest taps (or fewer, if less than 6 were larger than 0.002) were retained. These taps were equally spaced, with a spacing equal to 2*N_f samples at 8kHz audio sampling rate.
As before, the parameter A determines the magnitude of the amplitude variation and P determines the magnitude of the phase variation. N_f determines the period of both variations, and hence the spacing of the taps in the impulse response, which is the main factor determining the mean delay of signals passing through the filter. In fact, the mean delay is roughly equal to the time between the first tap and the largest tap of the filter impulse response. In the filters we have looked at the largest tap is either the 2nd or 3rd non-zero tap. This leads to a mean delay of roughly either 2*N_f samples, equivalent to N_f/4 or N_f/2ms, respectively. Different parameter combinations were tested. Typical plots of the filter responses are shown in Figure 8(a) to (d).
These filters were found to be successful. In general they give satisfactory single transmitter output as well as an improvement in two transmitter output over the whole range of differential delays tested. Some parameter combinations produced better results than others, although in some cases it was very hard to choose between them. Overall the flavoured parameter set was A=0.3, P=0.3, N_f=8.

Example 8

Sinusoidal amplitude variation over frequency for both transmitters.
This example considers the effects of only varying either amplitude over frequency, by setting P = 0 in the formula of Example 7 for the frequency response. This produced acceptable results although not as clear as Example 7.

Example 9

Sinusoidal phase variation over frequency for both transmitters.
By setting A to zero in the frequency response formula given for example 2, it was possible also to investigate the effect of varying phase only. For both one and two transmitters, the received signals were almost identical to the results for the preferred version of the amplitude and phase variation transmitter described as example 2. The best parameter set was that with P=0.3, N_f=8. Graphs of the filter response for this parameter set are given in Figures 9(a) and (b).
Since this arrangement does not significantly affect the amplitude of the transmitted signals, it was felt to be marginally preferable to Example 7. These filters were also suitable in that the mean delay they add to the signals was 4ms which is comparable to the fixed delay currently used.
As noted above, the degradation in the received signals when there is a differential delay occurs because the interference produces nulls in certain parts of the audio spectrum. In addition to the filtering discussed above it is possible to increase the power or pre-emphasise the most important parts of the spectrum so that the effect of the nulls is not so severe.
For example, it is possible to carry out pre-emphasis in conjunction with the sinusoidal phase-only filter of Example 9. Prior to passing through that filter, the transmitted signals can be filtered through a modified band pass filter. This may be a 1st order Butterworth filter, but with its coefficients adjusted so that the output of the filter is equal to the input plus a variable proportion of the original band-pass filter. Thus, if the frequency response of the unmodified 1st order Butterworth filter H(f) is given by: H(f) = B/A, where B and A are the filter coefficients, the frequency response, H(f) of the modified filter is given by H'(f) = 1+K(B/A), where K is a constant which determines the proportion of the original filter that is added to the original signal. The frequencies which might require pre-emphasis are in the 500-2000 Hz range.
There are a small number of air traffic control sectors which use three transmitters rather than two. For such situations although pre-filtering two out of the three transmitted signals may reduce distortion due to differential delay, the preferred solution is to pre-filter signals at all three transmitters.

Example 10

Combinations of sinusoidal phase and amplitude variation for three transmitters.
The filter for each transmitter was based on Example 7, but the three transmitters had different combinations of A and P. The frequency responses were given by: H 1 (f) = (1 - A 1 sin(2πNff'+π/2)) exp(-jπP 1 sin(2πNff')) H 2 (f) = (1 + A 2 sin(2πNff'+ π/2)) exp(+jπP 2 sin(2πNff')) H 3 (f) = (1 + A 3 sin(2πNff'+ π/2)) exp(-jπP 3 sin(2πNff')) where the parameters were chosen to be:
Tx1: A₁=0.3, P₁=0.3
Tx2: A₂=0.0, P₂=0.3
Tx3: A₃=0.3, P₃=0.0

		Tx1	Tx2	Tx3
1	Ampl.	1	1	1
	Delays (ms)	0	2	0-25
2	Ampl.	1	0	1
	Delays (ms)	0	-	0-25

3	Ampl.	1	1	0
	Delays (ms)	0	0-25	-
4	Ampl.	0	1	1
	Delays (ms)	-	0	0-25

NB: 2ms was chosen for the fixed differential delay for Tx2 in case 1 because this delay is one of the worst. Results for other delay values should be better, since the combined signal from Txland Tx2 would be degraded less.
Case 1 simulates situations where the aircraft is roughly equidistant from all 3 transmitters (this can only be true over a small region of space). Cases 2, 3 and 4 simulate situations where the aircraft is equidistant from two of the transmitters but quite far from the third.
The results for cases 1 and 3 were good, with the received signals remaining quite clear over the range of delays. The results for cases 2 and 4 were less satisfactory, with the received signals sounding quite thin and with noticeable degradation still present at low delays. This is consistent with poorer performance when one of the dominant transmitters is Tx3, which does not use phase variation.

Example 11

Sinusoidal variation in phase only, but with different values of N_f.
An other alternative is to use just a phase-only variation (since amplitude variation is less desirable in general), but with a different period for one of the transmitters. In this case, the frequency responses were given by: H 1 (f) = exp(-jπP sin(2πNf 1 f')) H 2 (f) = exp(-jπP sin(2πNf 2 f')) H 3 (f) = exp(+jπP sin(2πNf 2 f')) where P = 0.3, N_f1, = 7 and N_f2 = 8. Filter responses for the 3 filters are shown graphically in Figures 10(a) to (c).
The same combinations of signal amplitudes and delays as in section Example 10 were examined. (Case 4 is the same as the two transmitter case investigated in Example 9, and so the simulations were not repeated).
The results for cases 2 and 3 were good. The received signals sounded roughly similar over a wide range of delays and were almost as clear as for the best two-transmitter results. For case 1, the received signals did not sound as good - they were harsher in tone and there was a slight high pitched background whine. However, they were still better than the unprocessed signals. Finally, since there is no amplitude modulation being carried out, the single transmitter cases will also still be good.

FILTER COEFFICIENTS

The output of a filter is defined by its a and b coefficients as follows:
a ₀ y(n) = b ₀ x(n) +b ₁ x(n - 1) + ... + b_kx(n - k) + ... - a ₁ y(n - 1) - a ₂ y(n - 2) - ... - a_ky(n - k) - ..., where y(n) is the n ^th sample of the output and x(n) is the n ^th sample of the input.
The coefficients for the two recommended filters, and for the pre-emphasis filter, are listed below. Any coefficients not listed in the table are equal to zero.
Forth filter types in Examples 7 and 9, the non-zero b coefficients are equally spaced - the spacing, Δ, varies with the value of N_f' . For N_f= 6, 8 and 12, Δ takes the values 12, 16 and 24, respectively. The notation b_Δ means the Δth b-coefficient, i.e. for Δ =16, b_Δ, bo and b_2Δ are the 1st, 17th and 33rd coefficients.
For the three-transmitter case, two of the transmitters should have the same coefficients and spacings as given in the table for Example 9. The third should have the same coefficients as for Tx 1, but with a spacing of Δ =14 (equivalent to N_f = 7).

Preferred filters

Filter type	Parameters	Tx	Coefficients
EXAMPLE 7	A =0.0, P =0.3	1	a₀=1.0
	(3)	b₀=0.103, b_Δ=-0.421, b_2Δ=0.790, b_3Δ=0.421,
		b_4Δ=0.103, b_5Δ=0.017
		2	a₀=1.0
		b₀=0.103, b_Δ=0.421 b_2Δ=0.790, b_3Δ=-0.421, b_4Δ=0.103,
		b_5Δ=-0.017
EXAMPLE 9	A =0.3, P =0.3	1	a₀=1.0
EXAMPLE 9		b₀=0.169, b_Δ=-0.555, b_2Δ=0790, b_3Δ=0.287, b_4Δ=0.037
		2	a₀=1.0
		b₀=0.169, b_Δ=0.555, b_2Δ=0.790, b_3Δ=-0.287, b_4Δ=0.037

Pre-emphasis filter

	Parameters	Tx	Coefficients
	k= 1 1250-1750Hz	1, 2	a₀=1.0, a₁ =-0.651, a₂=0.668
b₀=1.166, b₁=-0.651, b₂=0.502

As noted above, the preferred filter for a two transmitter system is that of Example 9. Figures 11(a) and 12(a) show variations with time of a typical voice signal received from two transmitters separated by differential delays of 0.25 ms and 5.00 ms respectively. Figures 11(b) and 12(b) show power spectral density plots for transmitted and received signals, with the signals having been prefiltered according to the invention with the filter of Example 9. The results obtained in both cases are comparable with the received signal shown in Figure 3(b). Thus, the filter of Example 9 produces acceptable results over a range of differential delays. Therefore, with a filter of this type in place, differences in the differential delay caused by re-routing of the network 11 do not seriously degrade the signal.

Claims

A transmitter for use in an audio communications system having means for transmitting modulated audio signals to a communications network and means for processing audio signals prior to transmission whereby to change at least one parameter of the audio signal by an amount which is a function of time and/or the frequency of the audio signal.
A transmitter as claimed in claim 1 in which the amount of change of the parameter of the signals varies periodically with frequency.
A transmitter as claimed in claim 2 in which the amount of change of the parameter of the signals varies sinusoidally with frequency.
A transmitter as claimed in claim 1 in which the amount of change of the parameter of the signals varies periodically with time.
A transmitter as claimed in claim 4 in which the amount of change of the parameter of the signals varies sinusoidally with time.
A transmitter as claimed in claim 1 in which the amount of change of the parameter changes randomly with frequency and/or time.
A transmitter as claimed in any preceding claim in which the phase of the signals is changed.
A transmitter as claimed in any preceding claim in which the amplitude of the signals is changed.
An audio communications system including an audio signal source, at least two transmitters for transmitting identical audio signals generated by the audio signal source to one or more receivers, and network means for conveying signals from the audio signal source to the transmitters, wherein the transmitters transmit to the receiver(s) modulated signals which are demodulated by the receiver(s) and at least one of the transmitters is provided with means for processing signals received from the network means prior to transmission whereby to change at least one parameter of the received signals by an amount which is a function of time and/or the frequency of the received signal.
A system as claimed in claim 9 in which the amount of change of the parameter of the signals varies periodically with frequency.
A system as claimed in claim 10 in which the amount of change of the parameter of the signals varies sinusoidally with frequency.
A system as claimed in claim 9 in which the amount of change of the parameter of the signals varies periodically with time.
A system as claimed in claim 12 in which the amount of change of the parameter of the signals varies sinusoidally with time.
A system as claimed in claim 9 in which the amount of change of the parameter changes randomly with frequency and/or time.
A system as claimed in any of claims 9 to 14, wherein the phase of the signals is changed.
A system as claimed in any of claims 9 to 15 wherein the amplitude of the signals is changed.
A system as claimed in any of claims 9 to 16 including two transmitters in which only one of the transmitters is provided with said signal processing means.
A system as claimed in any of claims 9 to 16 including two transmitters in which both transmitters are provided with said signal processing means.
A system as claimed in claim 18 when dependent on claim 10 or 12 in which both transmitters are provided with signal processing means for changing the phase of the signal periodically with frequency or time and in which the variations of phase with frequency of the two signal processing means are 180° out of phase with each other.
A system as claimed in claim 18 or 19 in which the magnitude of change of the parameter is the same for both signal processing means.
A system as claimed in of claims 18 to 20 including three transmitters in which all three transmitters are provided with said signal processing means.
A system as claimed in claim 21 in which the three signal processing means each have different amplitude responses.
A system as claimed in claim 21 or 22 in which the three signal processing means change a parameter of the received signals by an amount which varies periodically with frequency and wherein the three signal processing means each have different periods.
A system as claimed in claim 21 or 22 in which the three signal processing means change the amplitude and phase of input signals and each have different combinations of parameters for amplitude and phase.
A system as claimed in any of claims 9 to 24 or a transmitter as claimed in any of claims 1 to 8 in which at least one transmitter further includes means for pre-emphasising certain frequencies prior to processing.
A system as claimed in any of claims 9 to 25 in which the transmitters are stationary.
A method of operating an audio communications system including an audio signal source, at least two transmitters for transmitting identical audio signals generated by the audio signal source to one or more receivers, and network means for conveying signals from the audio signal source to the transmitters, wherein the transmitters transmit to the receiver(s) modulated signals which are demodulated by the receiver(s), the method comprising processing, in at least one of the transmitters, processing signals received from the network means prior to transmission whereby to change at least one parameter of the received signals by an amount which is a function of time and/or the frequency of the received signal.
A transmitter substantially as hereinbefore described with reference to the accompanying drawings.
A communications system substantially as hereinbefore described with reference to the accompanying drawings.
A method of operating an audio communications system substantially as hereinbefore described with reference to the accompanying drawings.